Decoding notch signaling pathway: an updated insight into its role in colorectal cancer progression and therapy

Abstract

Colorectal cancer (CRC) represents a major global health challenge, marked by a high mortality rate and a strong association with aging. A key factor contributing to the progression of this malignancy is the overactivation of oncogenic signaling pathways. Notch signaling is a conversed pathway. This pathway utilizes a distinct mechanism that relies on direct cell-to-cell interactions for activation. Notch signaling ligands and receptors possess unique biological functions and distinct structures. Aberrant Notch signaling plays a pivotal role in CRC progression. It coordinates diverse oncogenic programs, including sustained proliferative signaling, invasive behavior, cancer stem cell maintenance, acquisition of therapeutic resistance, and the promotion of angiogenesis through increased microvascular density. The interplay between Notch signaling and other oncogenic pathways leads to the development of CRC. Designing therapeutic agents that target the Notch pathway is a feasible strategy to mitigate the oncogenic features associated with CRC progression. Furthermore, assessing the components of this pathway can serve as a diagnostic biomarker and offer insights into patient prognosis. Predicting the response to treatment is another clinical application of Notch evaluation. This review delves into the multi-layered roles of Notch signaling in CRC.

Introduction

Epidemiological data clearly indicate the substantial global burden of colorectal cancer (CRC). It ranks as the third most commonly diagnosed malignancy and the second leading cause of cancer-related mortality worldwide. Advances in screening and early detection have been linked to a decline in CRC-related mortality during the previous two decades. However, the overall incidence of CRC continue to rise, and it is estimated that the annual number of new cases of CRC will reach about 3 million in 2040 [1]. CRC encompasses malignancies arising from the colon and/or rectum. As CRC is widely believed to originate from precursor lesions such as hyperplastic polyps. Early detection of these lesions through imaging modalities, particularly colonoscopy, is of critical importance [2]. The incidence of CRC has risen dramatically in people > 50 years of age, with about 10% of CRC reported in people < 50 years of age [3]. Additionally, diet, lifestyle, and anomalies such as Crohn’s and inflammatory bowel disease (IBD) also increase the risk of CRC [2].

CRC treatment options are limited to conventional approaches such as surgery, chemotherapy drugs, radiotherapy, and targeted therapy. Discovering the unique molecular features of CRC and integrating them with immunotherapy techniques may contribute to the development of more reliable therapeutic approach [4]. One of the pathways that has been preserved through evolution and exerts a targeted influence throughout critical stages of cell life is Notch signaling. Activation of this pathway has several effects, including regulating the cell cycle, tissue homeostasis, and dictating the fate of cells. Notch signaling is extensively involved in the initiation and progression of CRC, in addition to its contribution to epithelial-mesenchymal transition (EMT), angiogenesis, and tumor proliferation. Therefore, a deeper understanding of the cellular and molecular features of CRC may facilitate the development and optimization of targeted therapeutic approaches [5]. Rather than viewing Notch as a linear oncogenic pathway, we propose a unifying conceptual framework in which Notch signaling functions as a master regulator of colorectal cancer cellular plasticity, orchestrating stemness, EMT, angiogenic reprogramming, tumor microenvironment (TME) adaptation, and therapeutic resistance (Fig. 1). Although numerous studies have examined individual aspects of Notch signaling in CRC, the existing literature remains fragmented, largely focusing on isolated mechanisms rather than providing an integrated explanatory framework. Consequently, a cohesive model is currently lacking. To ensure a comprehensive and up-to-date synthesis of current evidence, we performed a structured literature search across key databases, including PubMed, Scopus, and Web of Science, without language restrictions. In this review, we address this gap by proposing a unifying conceptual framework in which Notch signaling functions as a master regulator of CRC cellular plasticity, integrating these interconnected processes and highlighting their translational and therapeutic implications.

Fig. 1.

This schematic model illustrates Notch signaling as a central integrative regulator of CRC cellular plasticity. Notch orchestrates interconnected processes including cancer stemness, epithelial-mesenchymal transition (EMT), angiogenic reprogramming, and tumor microenvironment (TME) adaptation. The bidirectional interactions among these dimensions highlight the role of Notch as a master regulatory hub driving tumor progression and translational vulnerability in CRC. These interconnected processes collectively constitute CRC cellular plasticity. Ultimately, this plasticity underlies therapeutic resistance, depicted as the overarching functional consequence of Notch-driven tumor reprogramming

Notch signaling; a CRC-focused overview

Compared to other signaling pathways, Notch signaling operates in a proteolysis-dependent manner. Oncogenic or tumor-suppressive impacts of Notch downstream effectors are highly context dependent.

Unlike many signaling cascades that rely on soluble mediators, the Notch pathway operates via contact-dependent intercellular communication. Activation is initiated when the Notch receptor on one cell interacts with transmembrane ligands expressed on a neighboring cell. Following the receptor-ligand interaction, the Notch receptor undergoes a conformational shift that exposes the S2 cleavage site for ADAM10/17 [6]. In order to release Notch intracellular domain (NICD), γ-Secretase cleaves the second site (S3) [7]. This sequential proteolysis releases the NICD, which translocates to the nucleus and regulates transcription. In CRC cells, NICD-driven transcriptional programs are closely associated with oncogenic properties. Importantly, structural instability within regulatory domains (particularly the negative regulatory region (NRR) and heterodimerization (HD) domain) may facilitate ligand-independent activation or enhance cleavage susceptibility. Additionally, domain-specific changes within the ankyrin (ANK) repeats of NICD may influence recruitment of nuclear co-activators [8, 9]. These alterations, by driving sustained NICD signaling, may contribute to the establishment of a pathological state within colorectal tissues.

The γ-secretase complex (composed of Presenilin, Nicastrin, Aph-1, and Pen-2) has been pharmacologically targeted using γ-secretase inhibitors (GSIs) [10]. However, in CRC clinical investigations, GSIs have demonstrated dose-limiting gastrointestinal toxicity, largely due to the physiological requirement of Notch signaling in intestinal epithelial homeostasis. In rats, potent GSIs (DBZ and BZ) induced dose-dependent intestinal goblet cell metaplasia within 5 days [11]. Consistently, in a phase I study, γ-secretase inhibition was associated primarily with mild (grade 1–2) toxicities [12].

Increasing attention has also been directed toward targeting downstream components of the pathway. For example, CB-103 disrupts the Notch transcriptional complex and has shown an acceptable safety profile in early-phase clinical evaluation (phase I/IIa) [13].

Monoclonal antibody-based strategies are another area of development. Demcizumab, an anti-DLL4 antibody, has demonstrated anti-angiogenic effects in preclinical studies [14]. In contrast, brontictuzumab (anti-Notch1) and tarextumab (anti-Notch2/3) encountered limitations in CRC clinical trials due to toxicity or failure to meet primary endpoints [15]. These outcomes underscore the importance of biomarker-driven patient stratification based on Notch pathway activation status. Current trends also emphasize combination strategies, including the integration of Notch inhibitors with immune checkpoint blockade (e.g., anti-PD-1/PD-L1), conventional chemotherapy such as 5-FU, or targeted agents like anti-EGFR therapies to overcome resistance. In parallel, advanced technologies (such as CAR-T cells directed against Notch-dependent cancer stem cells and nanoparticle-mediated delivery of siRNA or CRISPR-Cas9 systems) are being explored to enable more selective pathway modulation. Despite ongoing challenges future progress is likely to rely on biomarker-guided approaches and more selective downstream targeting strategies [16–18].

This therapeutic challenge underscores the importance of developing more selective strategies, including NRR-targeting antibodies, ligand-specific blockade, or isoform-selective modulation, to improve therapeutic index while minimizing systemic adverse effects.

Structural determinants of notch signaling with relevance to CRC

There are four distinct Notch receptors (Notch1-4). While their structural framework is similar, functional divergence among receptors has distinct implications in CRC. Within the extracellular domain (NECD), the NRR is composed of LIN-12/Notch repeats (LNR) and the HD domain, maintains the receptor in an autoinhibited conformation [19, 20]. In CRC, structural destabilization of the NRR or HD may facilitate ligand-independent activation, enhancing NICD production and sustaining oncogenic transcriptional programs. In this context, monoclonal antibodies targeting both wild-type and mutant NRR variants have shown promise in effectively suppressing Notch1 signaling [21].

The NICD contains several conserved domains, including RAM, ANK, nuclear localization signals (NLS), a transcriptional activation domain (TAD), and a C-terminal PEST sequence (Fig. 2) [22]. The ANK domain mediates protein-protein interactions required for transcriptional complex assembly. Meanwhile, the PEST sequence governs proteasomal degradation of NICD; disruption of this regulatory region may prolong NICD stability and amplify Notch signaling in tumor cells [23]. Notably, structural differences such as the absence of TAD in Notch3 and Notch4 may contribute to receptor-specific transcriptional outputs in CRC [24, 25].

Fig. 2.

A schematic representation illustrating the structural variations among distinct Notch receptor classes. Abbreviations: NICD (Notch intracellular domain), NECD (Notch extracellular domain), TAD (transcriptional activation domain), LNR (LIN-12/Notch repeats), NRR (negative regulatory region), HD (heterodimerization domain), TMD (transmembrane domain), ANK (Ankyrin repeats), NLS (nuclear localization signals), RAM (RBP-Jκ-associated molecule), HD-N (N-terminal portion of the HD), HD-C (C-terminal portion of the HD)

Notch activation is initiated by membrane-bound ligands including DLL1, DLL3, DLL4, Jagged1, and Jagged2, all of which contain a conserved DSL domain required for receptor interaction. The structures of these ligands are summarized in Fig. 3 [26, 27]. In CRC, differential expression of DLL4 and Jagged1 has been associated with tumor angiogenesis and modulation of the TME [28, 29]. Ligand-specific engagement may therefore influence signaling intensity, therapeutic vulnerability, and resistance mechanisms. Collectively, these structural domains not only determine canonical receptor activation but also shape mutation-driven signaling, ligand selectivity, and pharmacologic targetability in CRC. A domain-oriented understanding of Notch architecture is therefore essential for the development of selective therapeutic strategies with reduced gastrointestinal toxicity.

Fig. 3.

A schematic representation illustrating the structural variations among distinct Notch ligands. Abbreviations: SP (signal peptide), DSL (Delta/Serrate/Lag-2), PDZ (postsynaptic density protein), CR (cysteine-rich), DOS (Delta and OSM-11-like proteins)

Notch signaling target gene

The presence of NICD in the nucleus necessitates network-like interactions for target gene expression. The most important stage is the formation of the double- and triple-membered complexes (NICD/CSL and NICD/CSL/MAML). Target genes are maintained in a repressed state in the absence of a Notch pathway activator, though CSL links to a co-repressor. The three most well-known co-repressors implicated in CSL are SHARP, SMRT, and NCoR. These co-repressors increase chromosomal condensation and prevent the transcription machine from accessing the target gene through the utilization of histone deacetylase (HDAC). SMRT and NCoR block the transcription of a broad of genes and do not have a specialised role for the Notch pathway [30, 31]. On the other hand, SHARP exhibits relative specificity for the Notch pathway. Following recruitment by CSL, SHARP interacts with SMRT and HDAC1 [32].

As the primary co-activators of the Notch pathway, three proteins are recognised: p300, CBP (CREB-binding protein), and MAML (Mastermind-like) proteins. Unlike co-repressors, p300/CBP serves as a transcriptional co-activator. By recruiting histone acetyltransferases, it induces an open chromatin configuration that facilitates transcriptional complex assembly at target loci [33, 34]. MAML has a high specificity for the Notch pathway. MAML provides a scaffolding platform essential for the recruitment of transcriptional co-activators. Through its N-terminal domain, it links to the NICD-CSL complex and facilitates assembly of the transcriptional activation machinery [35].

Most prominent target genes of the Notch pathway are HES1 (Hairy and Enhancer of Split-1), HES5, HES6, HES7, HEY1 (Hairy/enhancer-of-split related with YRPW motif protein 1), and HEY2. In addition, HEY3 and HES6 are two more direct targets of Notch. In a recent study, the link between elevated HES1 expression and adverse clinical outcomes in CRC patients was mechanistically attributed to HES1-driven activation of the STAT3-MMP14 axis [36]. Mechanistically, the recruitment of additional proteins like co-repressors (primarily Groucho/TLE) is linked to this inhibitory behaviour. The basic functions of these inhibitory proteins are in the regulation of cell differentiation and the maintenance of the undifferentiated cell population within the tissue [37, 38].

Among Notch’s other target genes is the transcription factor c-Myc. This transcription factor promotes the expression of key Notch pathway constituents. Consequently, it reinforces Notch signaling through a synergistic co-activation mechanism. Other biological effects of the Notch pathway include upregulation of Deltex1, Cyclin D1, PTEN, DUSP1, and p21 [39, 40]. The NRARP (Notch Regulated Ankyrin Repeat Protein) protein is upregulated by Notch, thereby exerting negative feedback on Notch signaling. This protein facilitates the degradation of NICD by the ubiquitination/proteasome system. In other words, NRARP prevents the continuous activity of the Notch pathway [41].

Notch in colon tissue development: Microbiota crosstalk

Notch signaling can be considered an essential part of the complex series of alterations that occur throughout the formation of colon tissue at the molecular and cellular levels. The endoderm layer develops the primitive gut tube during embryonic stages, which has a simple structure. The primitive gut tube is divided into three regions: the foregut, midgut, and hindgut. Although the role of Notch signaling in early segmentation and primary gut tube formation remains incompletely defined, substantial evidence supports its essential contribution to hindgut-to-colon specification and later stages of intestinal development [42–44]. The primary mechanism by which Notch signaling influences colon tissue development is through transcriptional regulation. It modulates the activity of key transcription factors and controls the spatial localization of Notch receptors/ligands. A portion of the Notch-guided cell differentiation process involves mapping out the colon’s topography. Notch guides the formation of crypt and villus structures by directing lineage specification and cellular positioning. Moreover, Notch signaling ensures proper generation and spatial distribution of functional epithelial cell types. It regulates the differentiation of intestinal stem cells into specialized lineages, including enterocytes, goblet cells, and enteroendocrine cells [45–47].

The intricate process of preserving the homeostasis of the epithelium layer involves various molecular dimensions. Recent studies have highlighted the intricate interplay between Notch signaling and epithelial homeostasis [48]. Notch impedes the differentiation of stem cells by increasing the molecular level of suppressor transcription factors such as HES1 and HEY1 [49]. Thus, Notch signaling maintains the balance between newly generated cells and those lost due to damage or aging by preserving the stem cell pool. Furthermore, the response of stem cells to various stimuli that are present in their microenvironment (niche) is mediated by the Notch [50]. This improves the colon tissue’s capacity for healing from damage by sustaining the balance between differentiated and undifferentiated cells. The Notch pathway facilitates the integration and regulation of the immune system in the colon. Notch enhances the expression of pro-inflammatory cytokines in response to damage in order to deal with the pathological circumstances [51]. Another facet of tissue homeostasis is the integrity of the epithelium layer. Notch signaling contributes to the structural integrity of the colonic epithelium by regulating the composition of extracellular matrix (ECM) components and modulating intercellular adhesion dynamics. Through these mechanisms, it reinforces epithelial barrier function and maintains tissue stability [52].

The bidirectional interaction between the gut microbiota and the Notch signaling pathway has emerged as a pivotal regulatory axis. Accumulating experimental and transcriptomic evidence supports a mechanistically and statistically robust crosstalk that is particularly relevant in intestinal inflammatory disorders and colorectal neoplasia. Notch activation promotes differentiation toward absorptive enterocytes while suppressing secretory lineages, including goblet and enteroendocrine cells. Microbial colonization critically modulates this balance. In germ-free and gnotobiotic zebrafish models, excessive Notch activity results in a marked depletion of secretory cells. Conversely, microbiota-derived signals attenuate Notch activation in a MyD88-dependent manner, thereby facilitating secretory lineage differentiation [53].

Importantly, this interaction is bidirectional. While microbial communities regulate epithelial Notch activity, Notch signaling itself influences microbial composition and inflammatory tone. Beneficial taxa such as Lactobacillus and Akkermansia muciniphila have been associated with reduced inflammatory cytokine production (e.g., IL-6 and TNF-α) and enrichment of short-chain fatty acid-producing bacteria, partly through modulation of Notch-dependent pathways [54]. In contrast, sulfate-reducing bacteria such as Desulfovibrio vulgaris have been shown to directly stimulate Notch signaling in a dose- and time-dependent manner. Exposure to D. vulgaris markedly upregulated Notch1 expression (~ 4-fold) and increased NICD accumulation (up to 6.7-fold). Concurrently, significant induction of SOCS3 (~ 13.7-fold) and pro-IL-1β (up to 29-fold) was observed in macrophages and HCT116 epithelial cells [55]. Pharmacologic inhibition with DAPT (a GSI) or Notch1 silencing attenuated these effects, underlining a causative link between sulfate-reducing bacterial exposure and Notch-driven inflammatory amplification.

Microbiome modulation has emerged as a potential indirect regulator of Notch signaling in CRC. Multi-strain Lactobacillus formulations suppress cell proliferation and apoptosis through downregulation of key Notch mediators and concurrent inhibition of Wnt/β-catenin signaling [56]. Similarly, Lactobacillus acidophilus attenuates Notch1 activity and improves mucosal integrity in inflammation-associated models [57]. Other commensals such as Akkermansia muciniphila and Bifidobacterium species may further dampen Notch-driven inflammation via short-chain fatty acid production. In contrast, antibiotics might reduce Notch-activating microbes, but their non-selective effects and limited mechanistic evidence restrict their therapeutic value. Overall, strain specificity, microbiome heterogeneity, and potential toxicity from excessive Notch inhibition necessitate further validation. Collectively, probiotic-based microbiome modulation represents a promising yet preliminary approach for indirect Notch targeting in CRC that requires robust clinical confirmation.

Within the intestinal stem cell niche, Notch signaling functionally integrates with the Wnt/β-catenin axis to sustain stem cell identity and orchestrate lineage allocation between absorptive enterocytes and secretory cell populations. Moreover, microbiota-derived metabolites and pattern-recognition receptor-mediated signaling dynamically interface with these pathways, thereby modulating epithelial turnover and barrier homeostasis [58]. Beyond epithelial differentiation, Notch contributes to barrier integrity through modulating of tight junction components and suppression of MLCK/p-MLC-mediated cytoskeletal contraction. Activation of the Jaggged1/Notch1/Hes1 axis has been associated with enhanced barrier protection, whereas pharmacologic inhibition (e.g., LY411,575) exacerbates experimental colitis [59].

Functional consequences of notch signaling in CRC

The uncontrolled activity of the Notch pathway can be attributed to multiple molecular mechanisms. One of the most well-known reasons for hyperactivity is mutation. Given the complexity of the genomic landscape, mutations may affect virtually any component of the Notch signaling cascade. However, studies have reported amplification in NOTCH1 in about 22% of CRC cases. Mutant NOTCH1 has altered HD and NECD domains, which induce conformational shifts that enhance receptor activation [60, 61].

In a large sequencing series of 1,154 mCRC, coding alterations in Notch receptors were identified with variable frequencies. Nonsynonymous variants occurred most often in NOTCH1 (11.5%), followed by NOTCH3 (10.4%) and NOTCH2 (4.4%), whereas NOTCH4 changes were uncommon. The predominant alteration type was missense, with mutations scattered rather than concentrated within specific functional hotspots. Truncating events were infrequent overall, although NOTCH3 showed a small subset (< 2%) enriched for such variants including a recurrent p.Asp847Ter/fs detected in 1.2% of tumors. No established activating hotspot mutations (such as NRR or PEST) were observed in this cohort [62].

In a pooled analysis of 4,341 CRC cases, the overall mutational architecture of NOTCH1, NOTCH2, NOTCH3, and NOTCH4 was largely comparable. For NOTCH1, presumed driver events were mainly truncating mutations, enriched within the extracellular region (~ aa 700–1200) and the intracellular segment (~ aa 1700–2100), which includes the ubiquitination site. A similar domain-oriented distribution was noted across the other Notch receptors [63].

In CRC, most Notch alterations are consistent with a loss-of-function phenotype and are linked to diminished pathway activity relative to wild-type tumors. Alterations in NOTCH3 impair receptor function and independently associate with favorable outcomes (HR 0.61). By contrast, Notch1 and Notch2 may display context-specific oncogenic behavior in defined molecular subsets, such as distal MSS tumors [62, 64].

Ligand overexpression is predominantly influenced by signals derived from neighboring cells rather than intrinsic alterations within the signal-receiving cell. The hyperactivity of signaling pathway may be linked to elevated stimulation. In the case of Notch, the increase in the ligands DLL4 and Jagged1 is more evident than other ligands in CRC [65, 66].

In addition to ligand amplification, the hyperactivity of the pathway is also explained by the rise of Notch receptors. The main root cause of receptor accumulation for this signaling pathway in the cell membrane is the inefficiency of the Notch receptor negative regulator system [67]. The total quantity of these signal receptors in the membrane is regulated by the endocytosis of Notch receptors. Conversely, endocytosis-dependent regulation prevents long-term stimulation or high sensitivity to ligands. Numb protein facilitates the endocytosis process of the Notch receptor. Molecular analyses have revealed functional insufficiency or downregulation of this protein in CRC [68, 69]. Of note, epigenetic modifications have the capacity to either downregulate the expression of negative regulators or raise the molecular level of the essential Notch pathway components. The simplest example is the alteration of gene expression caused by modifications to the promoters’ epigenetic patterns, such as hyper- or hypomethylation [70, 71].

In addition to the aforementioned, extrinsic factors within the TME, including stromal cells, inflammatory signals, and hypoxic stress, can significantly contribute to aberrant activation of the Notch signaling pathway. The stromal cells around the cell can provide signal induction by expressing the ligands of this pathway. Cancer-associated stromal fibroblasts (CAFs) are prominent in this regard because they overexpress the Jagged1 ligand. In this context, Varga J et al., proposed that the anti-invasive effect of Notch3 inhibition in tumor cells may be further enhanced when Notch3 signaling is simultaneously suppressed within the stromal compartment. This approach suggests a cooperative therapeutic benefit from dual targeting of tumor and stromal Notch3 activity [72]. Under hypoxic conditions, HIFs can directly interact with components of the Notch pathway. This interaction facilitates NICD stabilization and enhances its nuclear translocation. In this context, Wang H.-G. et al. demonstrated that disruption of the HIF-1α/Notch1 signaling cascade may substantially reduce the malignant behavior of CRC cells [73]. In chronic inflammatory conditions such as Crohn’s disease and ulcerative colitis, persistent exposure to pro-inflammatory cytokines, including TNF-α and interleukin-6, serves as a key trigger for Notch pathway activation [74]. Fazio C et al., demonstrated that pro-inflammatory stimuli markedly activate Notch1 signaling in CRC cells through MMP9 upregulation, thereby facilitating EMT and enhancing invasive capacity [75]. Additionally, inflammation-associated CRC is marked by early upregulation of Notch ligands (Jagged1/2) and receptors (Notch1/2) within the tumor-promoting microenvironment. However, downstream Notch target genes are not uniformly activated. These findings highlight context-dependent regulation during CRC progression [76].

To conclude, regulation of Notch signaling shows a distinct outcome in CRC. On one hand, its hyperactivation promotes tumorigenesis. On the other hand, predominant loss-of-function mutations may enhance anti-tumor immune cell infiltration and improve prognosis as well as responsiveness to immunotherapy. Therefore, a context-dependent understanding of Notch signaling is essential for designing targeted therapeutic strategies, including dual inhibition in both tumor and stromal compartments or leveraging loss-of-function mutations to optimize immunotherapeutic approaches.

Notch signaling and CRC stem cells

The features of cancer stem cells (CSCs) have positioned them as potential primary drivers and crucial contributors to cancer’s resistance. The most unique features of CSCs include stemness, pluripotency, the ability to remain in a quiescent state, and plasticity. Stemness, also referred to as the capability for asymmetric division in CSCs, includes the dual abilities of self-renewal and differentiation. Fundamentally, the proliferation of cancer cells and the maintenance of CSC pools are intrinsically linked to the concept of stemness [2]. CSCs possess the capacity to generate diverse cellular lineages with varying degrees of differentiation. This pluripotent potential represents a major driver of intratumoral heterogeneity [77]. Surprisingly, differentiated cells can return to an undifferentiated state under certain conditions, such as stress induced by chemotherapeutic agents. This phenomenon, known as plasticity, primarily enables CSCs to adapt to the microenvironment [78]. In addition to plasticity, another adaptive mechanism that CSCs employ to evade stress is the transient induction of a dormant state (quiescent) [79].

Growing evidence increasingly supports a central contribution of Notch signaling to the heterogeneity of CSC populations in CRC. Complementing this concept, a 2026 report revealed that Notch3 orchestrates pericyte phenotypic switching within the CRC microenvironment. Single-cell RNA sequencing analyses uncovered discrete pericyte subsets distinguished by differential Notch3 activity, where enhanced signaling was associated with a synthetic, proliferative state and compromised vascular integrity. Genetic ablation of Notch3 promoted vessel normalization and attenuated tumor growth [80]. Furthermore, recent single-cell transcriptomic profiling of primary CRC and matched liver and ovarian metastases has delineated a discrete stem-like epithelial population enriched for ASCL2 and PTPRO expression. Notably, this population was further subdivided into organ-biased subclusters. A DLL4-high subset was preferentially detected in primary tumors and demonstrated active engagement of the DLL4-Notch signaling axis in communication with endothelial cells and CAFs [81]. These data provide functional evidence that Notch signaling contributes to organ-specific metastatic dissemination by reinforcing stem-like transcriptional programs and facilitating CSC-microenvironment interactions. Rather than serving solely as a stemness-associated marker, Notch appears to stabilize metastasis-competent CSC states in a context-dependent manner.

Molecular-level studies have demonstrated a noteworthy resemblance between normal and cancerous stem cells, predominantly due to their shared key signaling pathways, such as the Notch pathway [82]. As previously stated, the primary functional mechanism of the Notch pathway involves inducing epigenetic modifications through the addition of methyl and acetyl groups. Notch orchestrates the regulation of methyltransferases, specifically by DNMT3A and DNMT3B, and acetyltransferase enzymes, which are pivotal in modulating this pathway [83]. Multiple functional investigations indicate that continuous Notch activity is essential for preserving CRC’s CSC properties. Disruption of this pathway diminishes self-renewal capacity and shifts cells toward secretory differentiation, accompanied by increased expression of lineage markers such as MUC2. Inhibition of receptor activation through ADAM17 blockade (e.g., MEDI3622 or TAPI-2) reduces NICD signaling and sphere formation, while targeting the ligand DLL4 similarly compromises CSC maintenance and tumor growth [84]. Together, these findings support a model in which ligand-driven Notch signaling sustains CSC stemness and prevents differentiation, thereby maintaining tumor-initiating potential.

It should be noted that Notch can exhibit either tumorigenic or antitumor properties, depending on the specific cellular and molecular context. Notch signaling sustains stem-like features by coordinating a broad CSC-associated transcriptional program. In CRC, increased PER3 expression attenuates Notch activity, leading to suppression of stemness-related properties [85, 86]. The Notch signaling suppresses differentiation by inducing transcriptional repressors like Hey1 and Hes1 [87]. Furthermore, Notch signaling inhibits apoptosis and blocks the induction of differentiation by downregulating p53 [88]. This axis can modulate cell differentiation and proliferation in a context-dependent manner by manipulating the interaction between the RB protein and the transcription factor E2F [89, 90]. On the other hand, research has demonstrated that heightened activity within this pathway correlates with the facilitation of asymmetric division (self-renewal). For instance, Notch3 levels are raised in response to elevated interleukin-6, subsequently enhancing the stem-like properties [91, 92].

CRC is thought to arise from CRC stem cells (Co-CSCs) located at the base of crypts. An examination of critical signaling pathways across three different cell line models (parental, colonospheres, and chemoresistant) revealed a predominant alteration in the Notch pathway, characterized by the upregulation of Notch1 and Hes1. Pharmacological inhibition of Notch using γ-secretase inhibitors (e.g., DAPT) significantly reduced cell survival by promoting apoptosis. This observation highlights the critical role of Notch signaling in maintaining the viability of colon cancer stem cells (Co-CSCs) [93, 94].

Notch1 is critical in regulating the stemness characteristics of Co-CSCs. Ongoing research is dedicated to identifying novel drugs to improve therapeutic outcomes in CRC. Research on α-mangostin has demonstrated that this compound suppresses stem-like properties by downregulating crucial components of the Notch signaling pathway, including Notch1, Hes1, and Hey1 [95, 96].

Collectively, current evidence positions Notch signaling as a context-specific integrative hub in CRC, coordinating both tumor cell-intrinsic and microenvironmental cues to sustain cellular diversity. By preserving stem-like transcriptional programs and limiting terminal differentiation, Notch activity reinforces CSC plasticity, metastatic adaptability, and therapeutic persistence. Single-cell transcriptomic analyses further support a model in which DLL4-Notch1/3 signaling within the stromal-vascular niche stabilizes discrete stem-competent states that fuel tumor evolution. Therefore, precise therapeutic modulation of Notch may disrupt the dynamic equilibrium that underlies CSC maintenance and heterogeneity-driven disease progression.

Notch signaling and CRC metastasis

CRC-related deaths are predominantly associated with metastatic progression rather than the primary tumor itself. Following local invasion, malignant cells disseminate systemically and preferentially colonize the liver and lungs. Metastasis can be conceptualized as a multi-step process encompassing various molecular interactions: the initiation of molecular changes, the entry into and survival within the bloodstream or lymphatic system, and the subsequent infiltration and colonization [97].

The Notch signaling regulate multiple stages of CRC metastasis. However, its influence is mainly pronounced during the survival phase within the circulatory system and the initial phase of metastasis [98, 99]. The presence of metastatic CRC cells within the circulatory system poses a noteworthy challenge to their survival. The activation of the Notch pathway confers resistance to these circulating tumor cells (CTCs) through a variety of mechanisms, thereby facilitating their capability to reach and establish at secondary sites. The primary impact of Notch signaling lies in its promotion of growth elements and inhibition of TRAIL-mediated apoptosis [97]. Notch activation facilitates cell cycle progression and the growth and colonization of CTCs by upregulating c-Myc and cyclin D1 expression [100–102]. Moreover, Notch signaling attenuates apoptosis through activation of the intrinsic pathway by upregulating anti-apoptotic mediators, including survivin and BCL-2 family proteins (BCL-2, BCL-XL, and MCL-1) [97]. This pathway further enhances tumor cell survival by promoting reactive oxygen species (ROS) detoxification and inducing autophagy. In parallel, it upregulates adhesion molecules, including integrins, facilitating platelet interactions that contribute to immune evasion [103–106].

Various processes can be involved in accelerating the initiation phase of metastasis; the EMT stands out as the most significant. EMT includes a wide set of molecular interactions that lead to the reduction of epithelial traits and the increase of mesenchymal characteristics in a reversible manner. The loss of epithelial cell-to-cell adhesions, exemplified by the downregulation of E-cadherin coupled with the upregulation of mesenchymal markers such as N-cadherin and vimentin, are pivotal molecular events driving EMT [107]. Furthermore, cytoskeletal remodeling, characterized by the formation of actin stress fibers, enhances the migratory capacity of tumor cells. Concurrently, increased production of matrix metalloproteinases (MMPs) promotes extracellular matrix (ECM) degradation, collectively driving invasive behavior and progression toward an aggressive metastatic phenotype [108]. In this regard, it has been documented that MMP-9 and MMP-2 upregulation can be induced by Notch1 [109]. EMT-related transcription factors (Snail, Slug, Twist, and Zeb1/2) are prominently expressed across the initiation stage of metastasis. The Notch pathway mediates EMT by establishing a transcriptional network focused on these regulators [97].

The impact of Notch signaling on EMT can be explored through the study of miRNA-598. Research has shown that miRNA-598 targeting the jagged1/Notch2 pathway leads to a decrease in EMT; consequently, the upregulation of this downregulated miRNA in CRC can be an appropriate approach to dealing with the aggressive CRC [110]. On the other hand, an increase in Notch1 activity leads to the activation of the Jagged1/Notch3 interaction, which subsequently induces EMT through upregulation of Slug and Snail [96, 111]. Not to mention that CRC invasiveness is amplified through promoting the mesenchymal phenotype, such as heightened cellular motility via decreasing E-cadherin and increasing N-cadherin and vimentin [112].

Hepatic metastasis affects 40–50% of CRC patients. In mouse models of serrated CRC harboring KrasG12D mutations, epithelial Notch1 activation markedly increases metastatic spread. Nearly all animals develop metastases, and over 80% show liver involvement, largely through TGF-β-driven neutrophil recruitment and TME remodeling [113]. Similarly, LIN28B contributes to CRC liver metastasis by inducing CLDN1 expression and activating Notch3 signaling. Clinically, high LIN28B expression was detected in 42% of primary tumors from patients who developed hepatic metastases and increased to 58% in matched metastatic lesions. Among these LIN28B-high metastases, 79% exhibited elevated CLDN1 and 63% showed increased Notch3 expression [114].

In mesenchymal CRC subtypes such as CMS4, AKT-mediated activation of Notch3 promotes liver metastasis. Consistently, a substantial fraction of human CRC liver metastases exhibits strong positivity for the Notch1 intracellular domain, reflecting active Notch signaling within metastatic lesions [115]. Over 22% of CRCs exhibit high copy number gains in Notch1, correlating with poor prognosis and EMT-driven metastasis [60]. In CRC, Notch receptor expression is elevated in approximately 29–73% of carcinomas and 11–56% of adenomas. Among them, Notch4 is overexpressed in nearly 56% of CRC cases and is significantly associated with liver metastasis compared with normal colonic tissue [116].

Within the liver niche, endothelial Notch signaling functions as a metastasis-restrictive pathway. Its activation attenuates ICAM1 expression on liver sinusoidal endothelial cells, thereby limiting tumor cell adhesion and subsequent seeding. Importantly, this antimetastatic effect appears to be independent of classical angiocrine mechanisms and has been shown to reduce metastatic burden in both CRC and melanoma experimental systems [117].

Direct evidence regarding Notch-Kupffer cell interactions in CRC remains limited. However, Notch signaling may facilitate immune evasion by sustaining inflammatory circuits and upregulating chemokines within the TME. Notably, alterations in the Notch pathway (reported in up to 69.2% of CMS1 tumors) have been associated with improved survival in CRC patients receiving immunotherapy [63].

Notch signaling and CRC angiogenesis

Angiogenesis refers to the formation of new blood vessels through sprouting from pre-existing vasculature. This process can also be defined as a mechanism leveraged by tumors to adapt to the hypoxic conditions prevailing in the TME [118]. HIF-1α and vascular endothelial growth factor (VEGF) are the core mediators of this process [119]. In contrast to normoxic conditions, hypoxia promotes the stabilization and accumulation of HIF-1α. This effect is mediated through the inactivation of prolyl hydroxylases (PHDs), which facilitates HIF-1α nuclear translocation and transcriptional activity [120]. The binding of this transcription factor to hypoxia-responsive elements (HREs) in the nucleus is facilitated by HIF-1β [121].

Typically, the target genes of this transcription factor have a pro-angiogenic nature. Accumulating evidence indicates that the HIF-1α/Notch axis operates in a bidirectionally reinforcing manner. Notch signaling enhances the stability and transcriptional activity of HIF-1α, while HIF-1α, in turn, potentiates Notch pathway activation [122, 123]. The promotion of VEGF expression by HIF-1α marks the first cellular step towards triggering angiogenesis. In response to hypoxia, tumor cells, by releasing VEGF, instigate the necessary cellular transformation for the onset of angiogenic sprouting [124]. In this regard, studies have identified hypoxia as associated with the hyperactivation of the HIF-1α/Notch/VEGF signaling axis [125].

This phenotypic shift underscores a previously underappreciated aspect of Notch-VEGF crosstalk in governing tip versus stalk cell fate determination. Such interaction guides the directional sprouting of nascent vessels along the VEGF gradient within the tumor microenvironment [126]. In other words, VEGF-induced specialization of endothelial cells (ECs) into leading-edge tip cells and proliferative stalk cells is further refined by Notch signaling. The cooperation between these pathways ensures precise regulation of sprout formation and vascular patterning. Notch, activated by VEGF, prevents destabilizing or excessive sprouting by inhibiting the differentiation of ECs into tip cells. Instead, it increases the expression of the stalk phenotype, thereby enabling the elongation of the formed sprout [28, 127, 128]. This phenomenon is based on the cell-to-cell signaling mechanism of Notch. VEGF leads to an increase in the Notch signal in adjacent cells by raising the membrane level of the ligand DLL4. DLL4, in turn, suppresses Notch signaling in VEGF-affected cells. Consequently, the tip cell population is regulated by VEGF signaling. Concurrent activation of Notch in adjacent stalk cells promotes the proliferation and growth necessary for sprout elongation [126, 129, 130]. Not to mention that the motility capability of tip cells is driven by a dynamic cytoskeleton and the formation of filopodia [131]. Two VEGF receptors (R1 and R2) have a known function in tip/stalk phenotype balancing. Stimulation of VEGFR2 leads to an increase in DLL4 levels, which promotes the tip cell phenotype. Conversely, VEGFR1, by binding to VEGF, controls the intensity of VEGF/VEGFR2 signaling and dominates the stalk phenotype [132].

Jagged1, a pro-angiogenic ligand present in stalk cells, leads to an increase in the number of tip cells through inhibiting the Notch/DLL4 signal [133]. Fringe glycosyltransferase enhance DLL4-mediate Notch signaling by modifying the Notch receptor has been documented. These enzymes boost the affinity of the Notch receptor of the stalk cells for DLL4, expressed in the tip cells. This process represents another mechanism for preventing uncontrolled amplification in the tip cells (Fig. 4) [134, 135].

Fig. 4.

Specification of endothelial cells to tip/stalk phenotype throughout sprouting angiogenesis. Tip cell; triggering the VEGF signal by VEGFR2 causes an increase in the expression of DLL4 and a decrease in Notch signaling levels. Fringe supports the low level of Notch in tip cells by enhancing DLL4-Notch binding. Jagged1 expressed in stalk cells leads to increased Notch signaling in tip cells and blocks the increase of DLL4. Sprout expansion towards high concentrations of VEGF. Stalk cell; stimulation of Notch signaling leads to an increase in proliferation and a reduction in the molecular level of DLL4 and VEGFR2. Increasing the expression of VEGFR1 has a negative effect on VEGFR2 and prevents the VEGF-induced tip cell phenotype. Abbreviations; VEGF (vascular endothelial growth factor), VEGFR1/2 (vascular endothelial growth factor receptor1/2), DLL4 (delta-like ligand 4)

Existing balance in the biological activity between the Jagged1 and DLL4 ligands promotes physiological angiogenesis. Induction of the ligand DLL4 through VEGF, which upregulates VEGFR2, leads to the orientation of the growing sprout towards hypoxic tumor cells by supporting the stalk phenotype. Downregulation of DLL4 accompanied by upregulation of Jagged1 and VEGFR1 favors tip cell specification. Consequently, aberrant sprouting from the originating vessel is amplified, leading to the development of a dense but functionally unstable vascular network (Fig. 5) [130, 136–139]. Increasing the migratory potential of ECs in the process of angiogenesis requires ECM remodeling by VEGF-induced Notch signaling. Research has clarified that applying DAPT to cells under low oxygen significantly reduces the expression levels of MMP2 and MMP9. However, in normoxia, blocking the Notch pathway does not notably impact the expression of MMP2 and MMP 9 [123, 140, 141].

Fig. 5.

DLL4/Jegged1 balance in angiogenesis and the opposite roles of VEGFR1 and VEGFR2 in the selection process of tip and stalk cells. A; Physiological angiogenesis during embryonic development. B; DLL4 blocking leads to a rise in the number of tip cells and subsequently increases the number of angiogenic branches and the instability of the vascular network. C; Suppression of Jagged1, together with an increase in LDL-4, leads to sprouting angiogenesis, associated with cancer. D; The Notch signal activation is linked with an increase in VEGFR1. E; VEGFR1 causes a shift to the stalk cell phenotype. F; VEGFR2 causes a shift to the tip cell phenotype. Abbreviations; VEGFR1/2 (vascular endothelial growth factor receptor 1/2), DLL4 (delta-like ligand 4)

Notch signaling and CRC drug resistance

Studies have revealed the intricate cooperation of multiple processes in conferring resistance to anticancer drugs. Improved DNA damage response (DDR), suppression of apoptosis, and promoting the active level of autophagy are among the most prominent drug resistance mechanisms. Moreover, cancer cells fortify resistance to treatment by modifying the pharmacokinetics of anticancer agents through upregulating efflux pumps and metabolic enzymes [2].

Further research is imperative to cover the existing knowledge gap regarding the impact of the Notch pathway on CRC drug resistance. Nevertheless, the role of Notch1 can be considered indisputable. Overall, the restoration of sensitivity in resistant CRCs can be strongly linked to the suppression of the Notch1/Hes1 axis. In this context, the molecular analysis of 121 CRC (stage II/III) samples with a 5-fluorouracil (5-FU) resistance phenotype unveiled an elevated level of Hes1 activity. The disruption or suppression of this inhibitory transcription factor resulted in a remarkable reduction in the expression of P-gp1 and ABCC1/2 [142]. The studies conducted on the effect of lncRNAs in drug resistance, particularly through modulation signaling pathways, are further confirmation of the resistance role of Notch1. Analysis of 108 CRC samples showed that 5-FU-resistant cells contained noticeable levels of LINC00152. In more detail, LINC00152 protects Notch1 activity by suppressing miR-139-5p. Targeting the LINC00152/miR-139-5p/Notch1 axis can be exploited to restore sensitivity to 5-FU and reduce the proliferation and invasiveness of CRC [143, 144]. However, these observations are largely based on retrospective analyses and relatively small cohorts. Whether the LINC00152/miR-139-5p/Notch1 axis independently predicts therapeutic resistance beyond established clinicopathological factors remains unclear.

The determination of the non-coding RNA (ncRNA) profile in CRC cells sheds light on the fact that downregulation of miR-195-5p coincides with a downregulation of P-gp and ABCG2 in 5-FU-resistant cells. Although the detailed mechanism of miR-195-5p in stemness and chemoresistance has not been fully identified, preliminary investigations have identified Notch2 and RBPJ as direct targets of miR-195-5p [145]. On the other hand, the increase of miR-34a is implicated in increasing the sensitivity to 5-FU by suppressing the DLL1/ABCG2 axis [146]. The stem cell-like phenotype characterized by the expression of key markers such as CD44 and CD133, has been implicated in 5-FU resistance. In this regard, an increase in miR-139-5p levels, which suppress Notch1, counteract CD44- and CD133-associated multidrug resistance [147]. It should be noted that reliance on CD44 and CD133 as definitive CSC markers remains controversial, as phenotypic plasticity and marker heterogeneity challenge the assumption of a stable stem-like subpopulation driving resistance.

The stem cell-like phenotype has also been blamed for inducing resistance to oxaliplatin, vincristine, and mitomycin C [147]. Targeted knockdown of Jagged2 as an adjunctive strategy in treatment with doxorubicin, oxaliplatin, and 5-FU leads to an increase in the active form of caspase 3. Thereby, it leads to an increase in the apoptotic removal of cells. However, the mechanism of Jagged2 in this resistance remains elusive [148]. Moreover, whether Jagged2-mediated resistance reflects ligand-specific canonical signaling or involves non-canonical Notch pathways has not been mechanistically dissected.

The DDR system encompasses several processes designed to avert genomic instability. Upregulation of DNA excision repair enzymes enhances the capacity to deal with treatment strategies based on DNA damage. Analyzing resistant cells to cisplatin and cytarabine (Ara-C) has reported blocking the positive correlation between Notch3 and key DNA excision repair enzymes (i.e., PARP1 and APE1) as an acceptable step in sensitization of resistant cells [149]. It remains uncertain whether Notch3 directly transcriptionally regulates PARP1 and APE1 or whether this association reflects broader stress-response network activation.

Resistance to targeted therapy can also be tied to Notch. Beyond the role of efflux pumps, angiogenesis can confer an effective mechanism for the drug clearance of anticancer drugs from the TME. Based on this principle, Notch 1 can precipitate the development of bevacizumab-resistant CRC by amplifying vascular perfusion [150, 151]. Undoubtedly, the direct effect of Notch signaling on drug-metabolizing enzymes in CRC remains inadequately explored; however, valuable discoveries are obtained in other malignancies. For instance, the 3- to 5-fold increase in ALDH levels in response to erlotinib in lung cancer cells employs a Notch3-dependent mechanism [10, 152].

Sensitivity to radiotherapy can also be manipulated by adjusting the Notch pathway. The upregulation of Notch1 is associated with an elevation in the levels of anti-apoptotic proteins such as XIAP, Bcl-xL, Bcl-2, and Mcl-1. The lethal nature of anticancer drugs and ionizing radiation can be contrasted with the survival/growth properties of Notch1 [153, 154]. In addition, Notch1 is the direct target of miR-665, and elevated levels of miR-665 are associated with an enhanced response to radiotherapy. Therefore, disrupting the circ-0005615/miR-665/Notch1 axis appears to be a promising strategy for therapeutic intervention [155]. In this line, the restoration of sensitivity to radiotherapy in resistant CRCs can be strongly linked to the suppression of the Notch1/Hes1 axis [156]. At first sight, targeting the Notch pathway appears to be a commendable strategy to deal with drug resistance in CRC; however, the available inhibitors are limited and have not received the necessary approval (Table 1).

Table 1.

Summary of the most important inhibitors of the Notch pathway

Inhibitor	Cancer	Mechanism of Action	FDA	Ref.
DAPT	Various (e.g., breast cancer)	γ-Secretase inhibitor	Not FDA-approved	[17]
RO4929097	Solid tumors	γ-Secretase inhibitor	Not FDA-approved	[159]
BMS-906,024	Leukemia	γ-Secretase inhibitor	Not FDA-approved	[160]
MK-0752	pancreatic cancer, Solid tumors	γ-Secretase inhibitor	Not FDA-approved	[161, 162]
Navicixizumab	Ovarian cancer, solid tumors	Monoclonal antibody; targets Notch ligands (DLL4)	Not FDA-approved	[163, 164]
OMP-59R5 (Tarextumab)	solid tumors	Monoclonal antibody; blocks Notch ligand DLL4 and Jagged1 binding to Notch2	Not FDA-approved	[165, 166]
AL101	Solid tumors, hematological malignancies	γ-Secretase inhibitor	Not FDA-approved	[167]
Avagacestat	Under investigation	γ-Secretase inhibitor	Not FDA-approved	[168, 169]
LY-411,575	Under investigation	γ-Secretase inhibitor	Not FDA-approved	[170]
Dibenzazepine (DBZ)	colorectal cancer cells, T-ALL	γ-Secretase inhibitor	Not FDA-approved	[170–172]
L-685,458	colorectal cancer cells, T-ALL	γ-Secretase inhibitor	Not FDA-approved	[171, 172]
TAPI-0	Melanoma, Renal Cell Carcinoma	ADAM inhibitor; suppress Adam17	Not FDA-approved	[173–175]
TMI-1	Breast cancer	ADAM inhibitor; suppress Adam17	Not FDA-approved	[176]

In a phase II single-arm trial of 37 heavily pretreated mCRC patients, the γ-secretase inhibitor RO4929097 (20 mg, 3 days on/4 days off) showed minimal activity: no objective responses were observed (ORR = 0), only six patients achieved stable disease, median progression-free survival was < 2 months, and median overall survival (OS) was 6 months. The investigators concluded that single-agent activity at this dose and schedule was very limited. Notably, despite evidence of elevated Notch pathway components (Notch1/ICN/HES1) in tumor tissues, this did not translate into clinical benefit [157]. Similarly, in a phase I study of 110 patients with advanced solid tumors (some CRC cases), crenigacestat (LY3039478) achieved marked Notch inhibition. However, clinical responses were rare and no meaningful overall response rate was observed. Gastrointestinal toxicity, thrombocytopenia, and colitis limited dosing [158]. Recent trials (2021–2024) indicate that limited efficacy of Notch-targeted therapies stems from three main factors: (i) dose-limiting gastrointestinal toxicity due to the pathway’s role in intestinal homeostasis. (ii) Lack of biomarker-based patient selection. (iii) Compensatory signaling mechanisms activation and Wnt crosstalk. Consequently, monotherapy has shown minimal benefit, underscoring the need for rational combination strategies. These clinical outcomes highlight a critical paradox: while preclinical models position Notch as a master regulator of resistance, systemic inhibition in patients produces limited efficacy and substantial toxicity. This discrepancy underscores the limitations of preclinical models in predicting therapeutic windows in human CRC.

Taken Together, Notch signaling acts as a central integrator of resistance in CRC, linking enhanced DNA damage response, apoptosis suppression, efflux pump upregulation, stem-like phenotype maintenance, autophagy, and angiogenesis into a coordinated resistance network. Its interaction with ncRNAs further stabilizes this phenotype, positioning Notch not merely as an associated factor but as a functional driver of chemoresistance and radioresistance. However, the pharmacological translation of this rationale remains limited. γ-secretase inhibitors are constrained by dose-limiting gastrointestinal toxicity due to non-selective blockade, monoclonal antibodies have shown modest efficacy, and robust predictive biomarkers are lacking. Thus, despite strong biological justification, effective clinical exploitation of Notch inhibition in CRC requires more selective, context-dependent, and combination-based therapeutic strategies.

Notch and the immune microenvironment

Recent evidence increasingly supports a pivotal role for Notch signaling in shaping the immune microenvironment of CRC. Early mechanistic studies demonstrated the dual immunomodulatory nature of Notch activity (Fig. 6).

Fig. 6.

Roles of Notch signaling in the colorectal tumor immune microenvironment. Tumor-intrinsic Notch1 and AKT-driven Notch3 activation promote transcriptional reprogramming, TGF-β–dependent neutrophil recruitment, myeloid expansion, and immune exclusion, thereby facilitating metastatic niche formation. Notch signaling also sustains PD-1 expression and T-cell exhaustion in CRC. Conversely, immune cell–intrinsic Notch signaling enhances CD8⁺ T-cell effector differentiation, supports Notch2/TGF-β-dependent TRM formation, and enables dendritic cell–mediated cross-priming. Spatial and genomic analyses further reveal mutation- and niche-dependent effects of Notch activity on immune infiltration and clinical outcome. Abbreviations: CRC (colorectal cancer), NICD (Notch intracellular domain), TAMs (tumor-associated macrophages), MDSCs (myeloid-derived suppressor cells), TRM (tissue-resident memory), MSI-H (microsatellite instability-high) CCR7 (C-C chemokine receptor type 7), PD-1 (programmed cell death protein 1)

Epithelial Notch signaling exerts non-cell-autonomous effects on the colorectal TME. In a pivotal study published in Cancer Cell (2019), Jackstadt et al. showed that constitutive activation of Notch1 in intestinal tumor cells transcriptionally reprograms the surrounding stroma and immune compartment. This shift drives tumors toward mesenchymal, poor-prognosis molecular subtypes. Mechanistically, epithelial Notch1 signaling enhances TGF-β-dependent neutrophil recruitment. This process promotes formation of a pro-metastatic niche and increases metastatic dissemination [113]. Whether neutrophil recruitment represents a driving factor or simply a factor associated with metastatic competence requires further elucidation.

Similarly, AKT-driven Notch3 activation was shown to sustain mesenchymal reprogramming and foster immune escape in CRC. In this regard, tumors with active Notch3 signaling exhibited diminished cytotoxic T-cell infiltration [72]. At the functional level, Notch3 activation was linked to expansion of immunosuppressive myeloid subsets, including TAMs and MDSCs, contributing to enhanced tumor growth in vivo [177]. These findings position Notch3 as a potential immunomodulatory target; however, systemic Notch inhibition has historically been limited by gastrointestinal toxicity.

Genomic analyses revealed that Notch-mutant tumors exhibited enhanced anti-tumor immunity, including increased PDCD1 expression and reduction in Tregs [178]. Additionally, loss-of-function mutations in the Notch pathway were associated with upregulation of chemokines such as CX3CL1 and CXCL1, alongside increased immune infiltration within the TME. Transcriptomic profiling revealed enrichment of both anti-tumor and pro-tumor immune subsets, indicating broad chemokine-driven immune recruitment [63]. However, transcriptomic profiling indicates enrichment of both anti- and pro-tumor immune subsets. This suggests that chemokine-driven recruitment does not uniformly translate into effective immune surveillance.

Subsequent investigations further refined the immunological landscape associated with Notch. Importantly, activation of the Wnt/Notch signaling axis was identified as a defining feature of an immune-low MSI-H subgroup [179]. This observation challenges the conventional assumption that MSI-H status uniformly predicts immunogenicity. Furthermore, this data suggests that intrinsic oncogenic signaling programs may override neoantigen-driven immune activation in a subset of tumors.

Beyond tumor-intrinsic effects, Notch directly regulates T-cell fate and functionality. Proprotein convertase-mediated proteolytic processing is essential for the proper activation of Notch receptors in T lymphocytes. This signaling axis sustains PD-1 expression and contributes to T-cell exhaustion in CRC. Disruption of this pathway attenuates inhibitory checkpoint signaling and reshapes the tumor immune microenvironment toward a more immunocompetent state. As a result, cytotoxic T-cell effector functions are enhanced, leading to suppression of tumor progression [180]. However, functional state of these T cells requires deeper single-cell and longitudinal validation.

In contrast to its exhaustion-promoting role, Notch signaling also contributes to the T cell differentiation. In this regard emerging data indicate that Notch signaling participates in the differentiation of tissue-resident memory (TRM) CD8⁺ T cells within the colorectal TME. Elevated expression of Notch2 in double-positive TRM cells, together with functional assays showing DLL4-TGF-β1-dependent enhancement of CD103 expression, support a cooperative interaction between Notch and TGF-β pathways. Epithelial Notch1 activation has been linked to stromal reprogramming and metastatic potentiation [181]. Noteworthy, Notch activity within CD8⁺ T cells appears to facilitate the generation of protective immune memory subsets. This functional divergence underscores the pleiotropic nature of Notch signaling.

Notch signaling further shapes anti-tumor immunity through dendritic cell lineage specification. In models of colitis-associated CRC, selective abrogation of Notch2 in dendritic cells disrupted cDC1 lineage commitment, attenuated CCR7-dependent migratory capacity, and impaired antigen cross-priming of CD8⁺ T cells. These defects facilitated inflammation-driven tumor progression. Conversely, restoration of Notch-competent dendritic cells re-established effective anti-tumor immunity, emphasizing the pivotal role of Notch2-dependent DC programming in constraining CRC [182].

Large-scale integrative analyses have further refined the immunological landscape associated with Notch signaling. Integrative TCGA analyses (n = 584) revealed that a Notch-related lncRNA risk signature stratifies CRC patients according to immune landscape features. Tumors with elevated NLncS scores displayed heightened immunogenicity, augmented CD8⁺ T-cell infiltration, and checkpoint pathway enrichment. This profile indicates potential sensitivity to CTLA4-targeted immunotherapy [183].

More recent high-resolution approaches have provided spatial and single-cell insights. Integrated single-cell and spatial transcriptome analyses of 27 CRC and matching liver metastases uncovered a metastatic niche enriched for MCAM⁺ fibroblasts. Computational inference suggested that these stromal cells may support the accumulation of CD8_CXCL13 T-cell subsets via Notch-mediated interactions. Notably, higher abundance of CD8_CXCL13 cells correlated with favorable patient survival outcomes [184]. Spatial transcriptomics further revealed Notch enrichment in poorly infiltrated metastatic niches. In this region coordinated activation of Notch and TGF-β pathways marked tumor cores with unfavorable prognosis [185].

Collectively, these findings establish Notch signaling as a bidirectional regulator of CRC immunity. Tumor-intrinsic Notch activation promotes immune exclusion, mesenchymal transition, and metastatic niche formation, whereas immune cell-intrinsic Notch signaling exerts both suppressive (T-cell exhaustion, myeloid expansion) and protective (TRM differentiation, cDC1 programming) effects. This functional duality highlights the necessity of isoform-, cell type-, and context-specific therapeutic strategies.

Crosstalk among notch signaling and other oncogenic pathways in CRC

Within signaling network, the Notch pathway assumes a key role in exacerbating the oncological features of CRC through interaction with other pathways. Therefore, concurrent inhibition of these interlinked pathways becomes imperative in the treatment of CRC as a complementary therapeutic approach [2, 97].

WNT/β-catenin signaling is considered a pivotal point in the onset of CRC and represents the most oncogenic linkage to CRC compared to other malignancies [186]. Notch-WNT pathway interaction has its own complexity, leading to the coinage of “Wntch” to encapsulate their entangled dynamics [187]. Applying the suppressive effect of Dishevelled on Notch and the β-catenin degradation complex (Notch-Off/WNT-On) and reciprocally inhibiting β-catenin by Notch (Notch-On/WNT-Off) underscores the dual states of interaction between these signaling cascades. The ability to quickly shift between these two modes is crucial for the fine-tuned modulation of target gene expression in both pathways, especially those implicated in determining cell fate [188, 189]. Generally, WNT-Notch interaction can be categorized into three principal facets: (i) the mutual upregulation of each pathway’s components by the other upon activation; (ii) co-operative regulation of target genes; (iii) the context-specific modulation of β-catenin by Notch and conversely [190, 191]. The WNT signaling cascade forms a positive regulatory loop by exerting an increasing effect on Notch2 and Jagged1 [192, 193]. It has been elucidated that the suppression of LGR5 leads to the down-regulation of the Notch pathway [194]. GSK-3β impedes proteasome-dependent degradation of Notch1’s NICD by phosphorylation [195]. Further evidence supporting the positive crosstalk between the Notch and Wnt pathways was provided by a study investigating PF-03084014, a γ-secretase inhibitor. Pharmacological inhibition led to a marked reduction in Notch expression, active β-catenin levels, and Axin2, a downstream Wnt-responsive gene [196].

PI3K/AKT and MAPK/ERK engage in pivotal oncogenic interactions with Notch [197]. One of the most important suppressors of the PI3K/AKT pathway is PTEN, which deactivates the pathway by converting the secondary messenger of PIP3 into PIP2 [198]. The Notch1 pathway triggers a downregulation of PTEN by upregulating Hes1. Enhancing AKT activity, in turn, provides the necessary stability through the phosphorylation of the Notch pathway [199, 200]. The close interplay between MAPK/ERK and Notch pathways can be manifested in the co-regulation of oncogenic gene expression. Increased metastatic potential, angiogenesis, and therapeutic resistance are prominent oncogenic traits linked to MAPK/ERK pathway activation. Importantly, MAPK/ERK signaling has been shown to positively regulate the expression of Jagged, DLL, and Notch family components [201, 202]. Evidence suggests that Notch1 modulates DUSPs, the primary regulators of the ERK pathway. The upregulation of Hes1 by Notch1 led to an elevation in the molecular level of phosphorylated ERK (active). In contrast, the depletion of HES1 in a shRNA-dependent manner resulted in a noticeable increase in the level of DUSPs and a concomitant decrease in the active form of ERK [203].

The delicate modulation of STAT3 (TGF-β) and NF-κB pathways leads to an increase in Notch activity, which is predominantly related to the upregulation of Notch1 and Jagged1 [204]. Further research has reported that the ablation of IκB kinase (IKK) results in suppressing the expression of Notch target genes [205].

The influence of prolactin on Notch1 activation in CRC is based on the induction of the JAK2-STAT3 and JAK2-ERK1/2 axes, culminating in the increase of Jagged1 expression [206]. Additionally, investigations into the interaction of the Hippo pathway with Notch have elucidated that the upregulation of Hes1 and Jagged1 is contingent upon the knockdown of Lats2 and Mst1/2 [207]. Collectively, a comprehensive investigation is required to unravel the interactive network of Notch. The oncogenic features previously reviewed, such as maintaining the stem-like state, metastasis, drug resistance, survival, and proliferation, are strengthened through the interaction of Notch with other pathways. These findings support the rationale for developing Notch-targeted therapeutic strategies.

Tumor marker and prognostic value of Notch signaling in CRC

The effectiveness of CRC treatment is profoundly dependent on early diagnosis. The accuracy of CRC diagnosis is tied to the dependability of biomarkers. In other words, CRC diagnosis is predominantly limited to a narrow range of biomarkers, such as carcinoembryonic antigen (CEA) and mutant forms of KRAS. Therefore, analysis of CRC will be accompanied by the discovery of unique molecular entities or biological patterns specific to the pathogenesis of CRC, refining precision in the diagnosis and optimizing therapeutic interventions in CRC [208, 209].

The aberrant activation of the Notch pathway can be regarded as an indicator for diagnosing and prognosticating CRC patients. Patients exhibiting Notch1^high compared to Notch1^low/− manifested markedly more aggressive CRC phenotypes [210, 211]. No clear association has been identified between elevated expression of Notch1 and age, gender, and tumor location; however, the escalation in the undifferentiated cell population that is correlated with the metastatic potential, particularly via the facilitation of EMT, can be attributed to Notch1. This escalation can be the main reason for the lower OS and 5-year survival rates of Notch1^high profile. Consequently, Notch1 can be defined as a prognostic marker independent of the known staging systems [212–215]. However, studies on the relationship between Notch1 and other clinical factors such as age, gender, and tumor location have shown inconsistent results across different cohorts, possibly due to differences in patient demographics and tumor subtypes. On the other hand, Notch2 is related to a more differentiated state of CRC. In prognostic evaluations, it stands at the opposite point of Notch1, which predicts a poor prognosis in Notch2^low /Notch1^high profile. This inverse correlation has been consistently reported in some cohorts, but in others, variations in tumor differentiation and clinical outcomes have led to mixed interpretations. The inverse correlation between Notch1 and Notch2 in CRC has been confirmed through analyzing 1,003 CRC samples. Notch2 absence reduced OS from favorable to approximately 45 months, marking it as a reliable predictor of death (HR = 1.69) [216, 217].

Analysis of 305 CRC specimens demonstrated nuclear expression of NOTCH1 (37%) and NOTCH3 (38%), both associated with tumor recurrence. In multivariate Cox modeling, nuclear Notch3^high independently predicted reduced distant relapse-free survival (HR = 1.71), with a markedly stronger impact in Stage II disease (HR = 3.47). Furthermore, nuclear Notch1^high and Notch3^high profile further exacerbated relapse risk (HR = 2.48) [218]. Notch3 leads to the development of malignancy by facilitating the infiltration of macrophages and myeloid-derived suppressor cells (MDSCs) into the TME. Therefore, targeting Notch3 could emerge as a therapeutic strategy to ameliorate the efficiency of the treatment [177, 219].

Further studies regarding Notch4 prognostic significance in CRC are necessary to reconcile the divergent results. Some studies have reported that high Notch4 expression correlates with suppression of proliferation, metastasis, relapse and stimulation of apoptosis in CRC [220]. Leveraging two pivotal public datasets (TCGA and GSE39582) and subsequent validation in an independent institutional cohort (n = 248), NOTCH4 expression was assessed. Its upregulation (particularly in liver metastases) was significantly associated with advanced nodal and metastatic stages, lymphovascular invasion, and CEA positivity; importantly, multivariate Cox analysis confirmed NOTCH4 as an independent predictor of inferior disease-free survival (DFS) (HR = 7.848) and OS (HR = 5.323) [221]. However, studies from other cohorts have reported contradictory findings, indicating a need for further validation in more diverse populations to reconcile these discrepancies.

It has been substantiated that DLL3 and DLL4 are linked with poor prognosis in CRC. Kaplan-Meier survival curves indicated a noteworthy survival disadvantage in patients with high DLL3^high profile. DLL3 levels were elevated in malignant versus adjacent tissues and further increased in metastatic lesions compared with primary tumors. Functional enrichment analyses suggested repression of AMPK signaling and mitophagy pathways in DLL3^high tumors, reinforcing its potential prognostic and diagnostic utility in colon adenocarcinoma. In a series of 327 CRC cases, loss of HES1 expression (64.2%) was linked to larger tumor size, lymphovascular invasion, and distant metastasis; whereas DLL4 overexpression (40.4%) was associated with perineural invasion, metastatic spread, and margin involvement. Multivariate survival analysis identified HES1 reduction (HR = 3.017) and DLL4 positivity (HR = 2.922) as independent predictors of worse OS. Interestingly, in certain patient groups, the expected link between these markers and negative outcomes was less pronounced. This inconsistency indicates that other molecular factors may be contributing to the observed results. Essentially, molecular interactions beyond DLL3 and DLL4 could be influencing tumor behavior. Furthermore, the combination of HES1^low/DLL4^high indicates the poorest prognosis [222–224]. Heightened Jagged1 expression in 80% of CRC cases has rationalized further studies. The metastatic nature of Jagged1, combined with the poor prognosis after surgery, has caused an elevated recurrence risk [225, 226]. Moreover, Jagged2 was identified in the majority of CRC specimens (~ 95%) and demonstrated nearly threefold higher expression in malignant tissue relative to adjacent normal mucosa. Functional suppression of Jagged2 significantly attenuated migratory and invasive behavior in CRC cell lines; nonetheless, definitive evidence regarding its independent prognostic significance is still lacking [227]. The prognostic evaluation of many antitumor drugs utilized in CRC management, such as checkpoint inhibitors (ICI), is profoundly impacted by the mutational status of the Notch signaling pathway. CRC cases harboring NOTCH alterations display a more immune-active TME, reflected by greater infiltration of CD8⁺ T cells, M1 macrophages, neutrophils, and activated NK cells. They are also characterized by higher tumor mutational burden, increased neoantigen load, and more frequent defects in DNA damage repair pathways. Multivariate Cox regression further demonstrated that NOTCH mutation status independently associates with improved outcomes following ICI therapy [228]. In a separate study employing the GSE108989 and TCGA datasets, it was demonstrated that mutations in the Notch signaling correlate with enhanced anti-tumor immunity in CRC, marked by higher CD8⁺ T cell infiltration and lower Treg cell levels. While these mutations did not influence OS or DFS, they were associated with earlier disease stages and reduced metastatic spread [178]. Silencing of NOTCH1 led to activation of chemokine-associated transcriptional programs, with marked elevation of CX3CL1 and CXCL1. Consistently, in ICI-treated CRC cohorts, tumors harboring NOTCH mutations were associated with significantly prolonged OS (multivariate Cox p = 0.006), supporting their role as predictive markers of immunotherapeutic benefit [63]. While some studies have demonstrated therapeutic benefits for CRC patients with Notch mutations treated with ICI, others have indicated that the efficacy may be influenced by the tumor’s specific molecular subtype and the patient’s immune response. These results further highlight the need for cohort-specific analysis.

Interpretation and comparison of findings

Context-dependency: The main cause of differences in results

A principal source of inconsistency across studies of Notch signaling lies in its highly context-dependent behavior (Fig. 7). In CRC, this functional plasticity is particularly apparent within CSC populations. Notch signaling sustains stemness and tumor plasticity by orchestrating epigenetic programs (e.g., DNMT3A/B and HATs). Concurrently, Notch represses differentiation through canonical effectors such as Hes1 and Hey1. scRNA-seq analyses of primary CRCs and matched metastatic lesions have further refined this model. Distinct CSC subpopulations characterized by DLL4^high profile have been identified. These cells engage in reciprocal crosstalk with endothelial cells and CAFs, thereby facilitating organotropic dissemination and promoting metastatic fitness.

Fig. 7.

Context-dependent roles of Notch signaling components in colorectal cancer. This figure summarizes the functions of Notch receptors (Notch1-4), ligands (DLL4, Jagged1/2), and the downstream effector HES1/HEY1-2 in regulating stemness, differentiation, EMT, metastasis, drug resistance, and angiogenesis. Additionally, the dual tumor-promoting and tumor-suppressive effects depending on the cellular and molecular context highlighted. Abbreviations: CRC (colorectal cancer), EMT (epithelial–mesenchymal transition), CSCs (cancer stem cells), OS (overall survival), DFS (disease-free survival), 5-FU (5-fluorouracil), P-gp (P-glycoprotein)

Conversely, emerging evidence supports context-specific inhibitory roles of Notch signaling. Upregulation of PER3 has been associated with attenuation of Notch activity and suppression of stem-like phenotypes. In parallel, genetic ablation of Notch3 in pericytes has been shown to induce vascular normalization and restrain tumor growth. Such discrepancies are largely attributable to differences in experimental frameworks. In vitro systems tend to emphasize the tumor-promoting role of Notch in CSC maintenance and frequently demonstrate sensitivity to γ-secretase inhibitors such as DAPT. In contrast, in vivo genetic models indicate that stromal-specific modulation of Notch signaling may yield anti-tumor outcomes.

Comparatively, in the metastatic cascade, the role of Notch in driving EMT appears relatively consistent across experimental systems. Notch activation promotes EMT (via Snail, Slug, and Twist), and enhances matrix remodeling (via MMPs). However, the survival advantage conferred to CTCs through anti-apoptotic mediators, including survivin and BCL-2, seems to be stage-dependent. Early-phase metastasis studies (e.g., miR-598 targeting of the Jagged1/Notch2) report attenuation of EMT. In contrast, more advanced models associate Notch1/Jagged1 signaling with enhanced invasiveness and metastatic competence. These discrepancies likely reflect intratumoral heterogeneity and temporal dynamics during disease progression. Primary tumors may preferentially exploit DLL4-Notch signaling to facilitate local invasion, whereas at metastatic sites such as the liver, Notch3 signaling appears to be reinforced to promote immune evasion and microenvironmental adaptation.

Findings in angiogenesis further illustrate this context specificity. The HIF-1α/Notch/VEGF axis functions bidirectionally to enhance hypoxic adaptation. Under physiological conditions, DLL4 regulates stalk cell elongation and vessel patterning. In tumor xenograft models an imbalance between Jagged1 and DLL4 results in aberrant sprouting and dysfunctional neovascularization. Hypoxia-focused experiments demonstrate more pronounced tumor-promoting effects than those observed under normoxia. This may explain why in vitro hypoxia models frequently report more aggressive phenotypes compared with tissue-based analyses.

These inconsistencies also observed in immune microenvironment. Depending on the cellular compartment, tumor stage, and its interaction with other signaling cascades, Notch can exert either tumor-promoting or tumor-restraining effects. For instance, in epithelial tumor cells, activation of Notch1 has been shown to promote TGF-β–mediated neutrophil recruitment, thereby fostering a pro-metastatic microenvironment and unfavorable prognosis. In contrast, within CD8⁺ T cells, Notch2 cooperates with TGF-β signaling to facilitate the differentiation of TRM subsets, strengthening protective anti-tumor immunity. This functional divergence highlights the non-cell-autonomous influence of Notch signaling on stromal and immune components of the tumor microenvironment.

Studies on therapeutic resistance reveal a similar degree of variability. Activation of the Notch1/Hes1 axis induces drug efflux transporters (e.g., P-gp and ABCC1/2) and promotes stem-like resistance to 5-FU and oxaliplatin in CRC cell lines. In contrast, early-phase clinical trials evaluating GSIs such as RO4929097 or crenigacestat have demonstrated limited efficacy. Gastrointestinal toxicity and the absence of robust predictive biomarkers have been major limiting factors. These discrepancies likely arise from differences between preclinical and clinical contexts. Preclinical systems often emphasize regulatory networks driven by ncRNAs (e.g., LINC00152/miR-139-5p/Notch1 axis), whereas in clinical settings, compensatory signaling pathways may attenuate the therapeutic impact of Notch inhibition.

Prognostic data are similarly heterogeneous. High Notch1 expression in large cohorts (e.g., n = 1,003) has been associated with poorer OS. However, correlations with age and sex remain inconsistent. These discrepancies may reflect molecular stratification, particularly differences between MSI-high and MSS tumors. Additionally, gene signatures involving Notch-related genes (e.g., Hey and Wnt5A) also predict prognosis and could guide stratification.

Methodological factors also contribute to divergent conclusions. Single-cell RNA sequencing and genetically engineered models, including gene deletion systems, often provide mechanistic insights that highlight context-specific or even anti-tumor functions of Notch. In contrast, expression-based studies, such as immunohistochemical analyses of clinical specimens, tend to emphasize tumor-promoting associations. Variability in cohort size, disease stage, and ethnic composition further complicates interpretation and may underlie conflicting observations, including the reported dual effects of Notch4 on proliferation and apoptosis.

The dual role of Notch signaling in CRC arises from receptor-specific functions, cellular/tumor heterogeneity, microenvironmental cues, and crosstalk with other pathways, rather than a uniform switch. The switch is highly context-dependent. Oncogenic dominance occurs in early-to-mid stage tumors with intact Wnt mutations and pro-inflammatory microenvironments (e.g., elevated Jagged1 ligand driving EMT). However suppressive roles emerge in advanced/metastatic settings with Notch pathway loss-of-function mutations, which paradoxically enhance anti-tumor immunity. Additionally, tumor heterogeneity (e.g., stem-like vs. differentiated cells) and epigenetic factors further modulate this balance.

These mechanistic insights enable actionable predictions: (1) High Notch1 expression or NOTCH1 gain predicts aggressive progression and potential benefit from selective Notch1 inhibitors or combination with Wnt modulators; (2) Notch pathway mutations (especially loss-of-function) serve as predictive biomarkers for enhanced immunotherapy response; (3) Integrated signatures (e.g., Notch1/Notch2 ratio or Hey/Wnt5A) can stratify patients for risk and guide therapy selection beyond descriptive associations. Thus, rather than a binary switch, Notch duality reflects adaptive signaling tuned by tumor context.

The integrative hub versus isolated effector paradigms

While single-effector models focus on individual downstream targets, the integrated node model conceptualizes Notch as a central regulatory hub embedded within TME-dependent networks and pathway crosstalk. This approach more effectively explains the convergence of tumorigenic processes and the maintenance of functional heterogeneity.

Within integrated framework, Notch coordinates several core oncogenic axes. First, by activating stemness-associated transcriptional programs, it sustains CSC heterogeneity. Second, through upregulation of survival and proliferation mediators such as c-Myc and cyclin D1, Notch functionally links EMT to enhanced metastatic competence. Third, by modulating ligand dynamics, including DLL4 and Jagged1, it regulates the phenotypic balance between tip and stalk endothelial cells during angiogenesis. In addition, Notch contributes to the organization of therapy-resistance networks by influencing DDR pathways, autophagy programs, and drug efflux transporter expression.

Evidence from scRNA-seq and multi-omics analyses supports “integrative node” model. These data suggest that Notch signaling stabilizes niche-dependent, metastasis-prone cellular states and correlates with adverse clinical outcomes. For example, nuclear Notch3 expression has been associated with reduced distant recurrence-free survival (HR = 1.71).

Single-effector models (such as defining Notch exclusively as a stemness marker or as an EMT driver) provide an incomplete interpretation. These frameworks neglect reciprocal signaling dynamics and context-specific regulatory feedback. As a result, they capture only a subset of Notch-dependent phenotypes.

The stemness-centered model is supported in chemoresistant CRC cell lines. For instance, α-mangostin suppresses Notch1/Hes1 activity and attenuates CSC-associated features. However, this paradigm does not explain tumor-suppressive contexts. PER3-dependent inhibition of tumor growth is one such example that falls outside a purely stemness-based interpretation.

Metastasis-focused models show similar limitations. They account for the survival and dissemination of CTCs. Yet, they do not adequately explain organotropism or the emergence of organ-specific metastatic subclusters shaped by local niche signals.

In contrast, the integrated node model is strengthened by pathway crosstalk. Notch amplifies WNT, PI3K/AKT, and MAPK/ERK pathways. These coordinated interactions generate a stable and self-reinforcing oncogenic network. This network-based structure better explains therapeutic persistence than linear, single-pathway models. Prognostic data further support this systems-level framework. Composite expression signatures (such as Notch1^high/Notch2^low or HES1^low/DLL4^high) outperform single biomarkers in risk stratification. Reported hazard ratios exceed 2.9 in some cohorts. Taken together, the integrated node model prioritizes functional network dynamics over isolated correlative observations.

This synthesis underscores that Notch’s oncogenic role is mechanistically contingent on tumor heterogeneity, microenvironmental cues, and pathway redundancy, explaining the limited efficacy of pan-Notch inhibitors (e.g., GSIs) in clinical trials. Future progress requires shifting from broad inhibition to precision strategies: receptor-specific antagonists, combination therapies targeting crosstalk nodes (e.g., Notch + Wnt or Notch + immune modulators), and biomarker-driven patient selection (e.g., Notch-mutations for immunotherapy synergy). Such a mechanistically grounded, predictive approach will be essential to translate Notch biology into clinically impactful interventions for CRC.

Conclusion and future directions

Tumor biology holds multifaceted and context-dependent significance across different stages of CRC, from initiation and development to metastasis and survival in adverse circumstances. A large body of evidence indicates that in CRC, the oncogenic nature of the Notch pathway predominates over the tumor suppressor properties, underscoring a considerable potential for therapeutic targeting. The existing limitations in CRC treatment options have made targeted modulation of Notch a promising strategy for enhancing personalized treatment plans. The complexity of the Notch pathway arises from its extensive crosstalk with other signaling networks, its genetic and epigenetic regulation, and its context-dependent activity within the TME. These multilayered interactions influence diverse biological processes, including apoptosis, autophagy, stem cell differentiation, and therapeutic resistance. Therefore, the advancement of potent therapeutics based on the Notch pathway to diversify CRC treatment options necessitates directing future research to clarify these complexities. Notch modulators. The evolution of the treatment perspective and the improvement of patient outcomes, aimed at diminishing CRC mortality, are linked to the depth of our understanding of the Notch pathway.

Although notable advances have expanded our understanding of Notch signaling in CRC, substantial mechanistic uncertainties persist. The differential contributions of individual Notch receptors and their corresponding ligands to tumor progression, metastatic dissemination and immune regulation remain insufficiently delineated. In addition, the molecular circuitry through which Notch activity confers resistance to frontline chemotherapeutic regimens requires more experimental dissection. The context-specific interplay between Notch signaling and diverse immune cell populations within the CRC ecosystem also remains only partially characterized.

Addressing these limitations will require integrative investigative strategies that transcend conventional bulk analyses. High-resolution single-cell and spatially resolved multi-omics platforms, together with patient-derived organoid systems incorporating immune components, offer promising avenues to capture the heterogeneity and spatial dynamics of Notch activation. Future clinical efforts should incorporate biomarker-guided patient stratification and evaluate isoform-selective or ligand-specific Notch modulation within rationally designed combination regimens.

Ultimately, effective clinical translation of Notch-targeted interventions in CRC will depend on coordinated molecular stratification, long-term clinical datasets, and integrative computational approaches that inform precision oncology.

Author contributions

Mehran Molavand, Azita Asadi, and Maryam Majidinia, wrote the article, and prepared the figures and tables; Bahman Yousefi, designed and revised the article. All the authors studied and approved the final manuscript. The authors declare that all data were generated in-house and that no paper mill was used.

AI Disclosure statement

The authors declare that ChatGPT (OpenAI, San Francisco, CA, USA) was used solely for language editing and improving the readability of the manuscript. The tool was not involved in the generation of scientific content, data analysis, or interpretation of results.

Funding

Not applicable.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.